Over the last several years, developments in peptide synthesis technology have resulted in automated synthesis of peptides accomplished through the use of solid phase synthesis methods. The solid phase synthesis chemistry that made this technology possible was first described in Merrifield et al. J. Amer. Chem. Soc., 85:2149-2154 (1963). The "Merrifield method" has for the most part remained largely unchanged and is used in nearly all automated peptide synthesizers available today.
In brief, the Merrifield method involves synthesis of a peptide chain on solid support resin particles. These particles typically consist of polystyrene cross-linked with divinyl benzene to form porous beads which are insoluble in both water and various organic solvents used in the synthesis protocol. The resin particles contain a fixed amount of amino- or hydroxylmethyl aromatic moiety which serves as the linkage point for the first amino acid in the peptide.
Attachment of the first amino acid entails chemically reacting its carboxyl-terminal (C-terminal) end with derivatized resin to form the carboxyl-terminal end of the oligopeptide. The alpha-amino end of the amino acid is typically blocked with a t-butoxy-carbonyl group (t-Boc) or with a 9-fluorenylmethyloxycarbonyl (Fmoc) group to prevent the amino group which could otherwise react from participating in the coupling reaction. The side chain groups of the amino acids, if reactive, are also blocked (or protected) by various benzyl-derived protecting groups in the form of ethers, thioethers, esters, and carbamates.
The next step and subsequent repetitive cycles involve deblocking the amino-terminal (N-terminal) resin-bound amino acid (or terminal residue of the peptide chain) to remove the alpha-amino blocking group, followed by chemical addition (coupling) of the next blocked amino acid. This process is repeated for however many cycles are necessary to synthesize the entire peptide chain of interest. After each of the coupling and deblocking steps, the resin-bound peptide is thoroughly washed to remove any residual reactants before proceeding to the next. The solid support particles facilitate removal of reagents at any given step as the resin and resin-bound peptide can be readily filtered and washed while being held in a column or device with porous openings.
Synthesized peptides are released from the resin by acid catalysis (typically with hydrofluoric acid or trifluoroacetic acid), which cleaves the peptide from the resin leaving an amide or carboxyl group on its C-terminal amino acid. Acidolytic cleavage also serves to remove the protecting groups from the side chains of the amino acids in the synthesized peptide. Finished peptides can then be purified by any one of a variety of chromatography methods.
Though most peptides are synthesized with the above described procedure using automated instruments, a recent advance in the solid phase method by R. A. Houghten allows for synthesis of multiple independent peptides simultaneously through manually performed means. The "Simultaneous Multiple Peptide Synthesis" ("SMPS") process is described in U.S. Pat. No. 4,631,211 (1986); Houghten, Proc. Natl. Acad. Sci., 82:5131-5135 (1985); Houghten et al., Int. J. Peptide Protein Res., 27:673-678 (1986); Houghten et al., Biotechniques, 4, 6, 522-528 (1986), and Houghten, U.S. Pat. No. 4,631,211, whose disclosures are incorporated by reference.
Illustratively, the SMPS process employs porous containers such as plastic bags to hold the solid support synthesis resin. A Merrifield-type solid-phase procedure is carried out with the resin-containing bags grouped together appropriately at any given step for addition of the same, desired amino acid residue. The bags are then washed, separated and regrouped for addition of subsequent same or different amino acid residues until peptides of the intended length and sequence have been synthesized on the separate resins within each respective bag.
That method allows multiple, but separate, peptides to be synthesized at one time, since the peptide-linked resins are maintained in their separate bags throughout the process. The SMPS method has been used to synthesize as many as 200 separate peptides by a single technician in as little as two weeks, a rate vastly exceeding the output of most automated peptide synthesizers.
A robotic device for automated multiple peptide synthesis has been recently commercialized. The device performs the sequential steps of multiple, separate solid phase peptide synthesis through iterative mechanical-intensive means. This instrument can synthesize up to 96 separate peptides at one time, but is limited at present by the quantity of its peptide yield.
Of interest is work by Geysen et al., which deals with methods for synthesizing peptides with specific sequences of amino acids and then using those peptides to identify reactions with various receptors. Geysen et al.'s work presupposes that one has a prior knowledge of the general nature of the sequences required for the particular receptors, so that the appropriate group of peptides can be synthesized. See U.S. Pat. Nos. 4,708,871 and 4,833,092; P.C.T. Publications Nos. WO 84/03506 and WO 84/03564; Geysen et al., Proc. Natl. Acad. Sci. U.S.A., 81:3998-4002 (1984); Geysen et al., Proc. Natl. Acad. Sci. U.S.A., 82:178-182 (1985); Geysen et al., in Synthetic Peptides as Antigens, 130-149 (1986); Geysen et al., J. Immunol. Meth., 1.02:259-274 (1987); and Schoofs et al., J. Immunol., 140:611-616 (1988).
Several research groups have reported the synthesis of synthetic combinatorial libraries of peptides. Those reports are discussed below.
In published PCT application PCT/AU85/00165 (WO 86/00991), Geysen describes a method for determining so-called "mimotopes". A mimotope is defined as a catamer (a polymer of precisely defined sequence formed by the condensation of a precise number of small molecules), which in at least one of its conformations has a surface region with the equivalent molecule topology to the epitope of which it is a mimic. An epitope is defined as the surface of an antigenic molecule which is delineated by the area of interaction with an antibody molecule.
The mimotopes are synthesized on a series of solid polymer (e.g. polyethylene with a coating of grafted acrylic acid) rods having a diameter of about 4 mm and a length of about 50 mm. A spacer formed by reaction of the .epsilon.-amino group of t-Boc-lysine methyl ester and then t-Boc-alanine was added to the resins, followed by removal of the t-Boc group to provide an amino group to be used to begin the syntheses.
A mixture of blocked amino acids containing different amounts of each of the blocked twenty amino acids to be used was dissolved in dimethyl formamide and then coupled to the rods. That first coupling was repeated three times using conventional solid phase synthesis techniques. Twenty amino acid residues were individually next added so that twenty 5-mer sequences were prepared, each having a single, known amino acid residue at the amino-terminus and a mixture of amino acid residues at each of the four other positions of the chain. Each of those twenty rod-linked peptides was then individually reacted with each of the twenty amino acid residues to form 400 (20.times.20) 6-mer peptides having the two amino-terminal positions defined and the four remaining positions as mixtures. Two more positions of mixtures of amino acids were then added, and the terminal amine acetylated to form N-acetyl 8-mers linked to the rods whose first two amino acid positions were undefined (mixtures), followed by two defined positions, followed by four undefined positions (mixtures), followed by the spacer and then the supporting rods.
The 400 rod-linked N-acetyl 8-mer peptide mixture preparations were then screened in an ELISA assay using a monoclonal antibody to a desired antigenic protein. The 8-mers having the best binding to the antibody were identified. Two sets of further 8-mers that contained the identified best-binding 2-mer sequences within those 8-mers were prepared.
A first set contained mixed amino acids at the three C-terminal positions, followed toward the N-terminus, by a position containing each of the twenty amino acids made by twenty separate couplings, the identified 2-mer sequences, two further mixtures at the next two positions, and an N-terminal acetyl group. The second group contained mixed amino acids at the four C-terminal positions, the identified 2-mer sequences, a position made by separate couplings of each of the twenty amino acids, mixed amino acids as the terminal residues and an N-terminal acetyl group.
Each of those rod-linked N-acetyl 8-mers was again screened in an ELISA with the monoclonal antibody. The best binding sequences for each group were identified, and thus 4-mer, best-binding sequences were identified.
The above process of separately adding each of the amino acids on either side of identified best-binding sequences was repeated until an optimum binding sequence was identified.
The above method, while elegant, suffers from several disadvantages. First, owing to the small size of each rod used, relatively small amounts of each peptide is produced. Second, each assay is carried out using the rod-linked peptides, rather than the free peptides in solution. Third, even though specific amounts of each blocked amino acid are used to prepare the mixed amino acid residues at the desired positions, there is no way of ascertaining that an equimolar amount of each residue is truly present at those positions.
In addition, Furka et al., (1988, 14th International Congress of Biochemistry, Volume 5, Abstract FR:013) described the synthesis of nine tetrapeptides each of which contained a single residue at each of the amino- and carboxy-termini and mixtures of three residues at each position therebetween. The abstract further asserts that those authors' experiments indicated that a mixture containing up to 180 pentapeptides could be easily synthesized in a single run. No biological assays were reported. More recently, Furka et al., Int. J. Peptide Protein Res., 37:487-493 (1991) reported on the synthesis of mixtures of 27 tetrapeptides and 180 pentapeptides prepared by physically mixing reacted resin-linked peptides. Those peptides were synthesized with one or mixtures of three or four residues at each position along the chain. No biological results using those relatively simple mixtures were reported.
Recent reports (Devlin et al., Science, 249:404-405 [1990] and Scott et al., Science, 249:386-390 [1990]) have described the use of recombinant DNA and bacterial expression to create highly complex mixtures of peptides. For example, a 45-nucleotide base pair stretch of DNA was synthesized in which the individual nucleotide bases were varied to contain all four possible nucleotide bases (guanine, adenine, cytosine and thymidine) at every position in the synthesized DNA chain, except at each third position (3, 6, 9, etc.) which contained only guanine and cytosine. The omission of adenine and thymidine at every third position in the synthesized DNA removed the possibility of chain terminator triplet codons ending in A or T, such as TAA or TGA.
The resulting DNA sequence would then code for a mixture of 15-mer peptides with all combinations of the 20 naturally occurring amino acids at each position.
Those investigators fused the 45 synthetic nucleotide sequence to a gene coding for the coat protein of a simple bacteriophage and created a large library of these bacteriophages. Each member of the library contained a different 45-mer DNA fusion sequence and therefore each member of the library resulted in a different 15-mer peptide fused to the outer coat protein of its corresponding fully assembled bacteriophage particle. Screening of the recombinant bacteriophage particles in a biochemical assay allowed the investigators to find individual peptide-coat protein fusions (bacteriophages) that were active in that assay by enrichment, selection and clonal isolation of the enriched bacteriophages that contained active peptide fusions. By determining the DNA sequence of the cloned bacteriophages, the investigators could deduce which peptide sequences were active in their assay.
That method yielded several peptide sequences from a mixture of 10.sup.7 or more recombinant bacteriophages. Each of the 15-mer peptides found contained the same four-amino-acid sequence somewhere within its overall sequence, thereby allegedly validating the assay accuracy and methodological approach.
The recombinant DNA method is extremely powerful for screening large numbers of peptides. However, it is limited in that the peptides must be fused to a larger protein as a result of and integral to the design of the method. The peptide-protein fusions (and corresponding bacteriophage particles) are likely to be unreactive in many biochemical, biological and in vivo assays where the peptides must be present in solution without steric hindrance or conformational distortion. In addition, the method results in an over-representation of some sequences of peptides due to the inherent redundancy of the genetic code which has several codons per amino acid in some cases and only one codon per amino acid in others.
Still further, neither group reported data as being definitive for the determination of optimal peptide ligands for strepavidin (Devlin et al.), or for the two monoclonal antibodies raised against myohemerythrin (Scott et al.). Neither group provided a single specific answer comparable to the expected sequence.
More recently, Fodor et al., Science, 251:767-773 (1991), described the solid phase synthesis of thousands of peptides or nucleotides on glass microscope slides treated with aminopropyltriethoxysilane to provide amine functional groups. Predetermined amino acids were then coupled to predefined areas of the slides by the use of photomasks. The photolabile protecting group NVOC (nitroveratryloxycarbonyl) was used as the amino-terminal protecting group.
By using irradiation, a photolabile protecting group and masking, an array of 1024 different peptides coupled to the slide was prepared in ten steps. Immunoreaction with a fluorescent-labeled monoclonal antibody was assayed with epifluorescence microscopy.
This elegant method is also limited by the small amount of peptide or oligonucleotide produced, by use of the synthesized peptide or nucleotide affixed to the slide, and also by the resolution of the photomasks. This method is also less useful where the epitope bound by the antibody is unknown because all of the possible sequences are not prepared.
The primary limitation of the above new approaches for the circumvention of individual screening of millions of individual peptides by the use of a combinatorial library is the inability of the peptides generated in those systems to interact in a "normal" manner with acceptor sites, analogous to natural interaction processes (i.e., in solution at a concentration relevant to the receptors, antibody binding sites, enzyme binding pockets, or the like being studied without the exclusion of a large percentage of the possible combinatorial library). Secondarily, the expression vector systems do not readily permit the incorporation of the D-forms of the natural amino acids or the wide variety of unnatural amine acids which would be of interest in the study or development of such interactions.
Yet another approach was reported by Lam et al., Letters to Nature, 354:82-84 (1991). Those workers reported the preparation of millions of bead-linked peptides, each bead containing a different peptide. The peptide-linked beads were reacted with a fluorescent- or enzyme-labeled acceptor, the beads bound by the acceptor were physically removed, and the sequence of the bound peptide was analyzed.
The interest in obtaining biologically active peptides for pharmaceutical, diagnostic and other uses would make desirable a procedure designed to find a mixture of peptides or a single peptide within a mixture with optimal activity for a target application. Screening mixtures of peptides enables the researcher to greatly simplify the search for useful therapeutic or diagnostic peptide compounds. Mixtures containing hundreds of thousands or more peptides should be readily screened since many biochemical, biological and small animal assays are sensitive enough to detect activity of compounds that have been diluted down to the nanogram or even picogram per milliliter range, the concentration range at which naturally occurring biological signals such as peptides and proteins operate.
Almost all of the broad diversity of biologically relevant ligand-receptor (or affector-acceptor) interactions occur in the presence of a complex milieu of other substances (i.e., proteins make up approximately 5-10 percent of plasma, e.g. albumin 1-3 percent, antibodies 2-5 percent-salts, lipids/fats, etc.). This is true for virtually all biologically active compounds since most are commonly present, and active, at nanomolar and lower concentrations. These compounds are also, in most instances, produced distant from their affection sites.
That a small peptide (or other molecule) can readily "find" an acceptor system, bind to it, and affect a necessary biological function prior to being cleared from the circulation or degraded suggested that a single specific peptide sequence can be present in a very wide diversity, and concentration, of other individual peptides and still be recognized by its particular acceptor system (antibody, cellular receptor, etc.). If one could devise a means to prepare and screen a synthetic combinatorial library of peptides, then the normal exquisite selectivity of biological affector/acceptor systems could be used to screen through vast numbers of synthetic oligopeptides.
The availability of a wide variety of clearly identified peptides in relatively limited mixtures would greatly facilitate the search for optimal peptides for any particular therapeutic end use application. At the present time, researchers are hampered by the inability to rapidly create, identify and screen large numbers of peptides with specific receptors. Work such as reported by Geysen has been valuable where the general nature of the required amino acid residue sequence could be previously determined, so that the specific peptides of interest could be individually formulated. However, such techniques cannot insure that the optimum peptides are identified for testing.
It would therefore be of considerable interest to have a method for the synthesis of mixtures of peptides in which individual residue positions can be specifically defined, such that a comprehensive array of peptides is available to researchers for the identification of one or more of the optimal peptides for reaction with receptors (acceptors) of interest, from which one can derive optimum therapeutic materials for treatment of various organism dysfunctions.